Scanning Tunneling Microscopy Investigation of Sulfide and

Vicki H. Wysocki , Jeanne E. Pemberton , T. Randall Lee , Ronald J. Wysocki , and Neal R. ... A. Shaporenko, P. Cyganik, M. Buck, A. Ulman, and M...
0 downloads 0 Views 2MB Size
Langmuir 1995,11, 506-511

506

Scanning Tunneling Microscopy Investigation of Sulfide and Alkanethiolate Adlayers on Ag(111) R. Heinzt and J. P. Rabe*f$ Max-Planck-Institut fur Polymerforschung, Postfach 3148,0-55021Mainz, Germany, and Institut fur Physikalische Chemie, Johannes-Gutenberg Universitat, Jakob-Welder-Weg 11, 0-55099Mainz, Germany Received May 16, 1994. I n Final Form: September 21, 1994@ Chemisorption of sulfides and alkanethiols on Ag(ll1) is studied by scanning tunneling microscopy (STM). Adsorption of sulfide from aqueous solutions of NazS and H2S or from gas phase with H S, respectively, leads to a highly ordered hexagonal adsorbate lattice (a1 = 440 f 10 pm) with a (d7x R10.9" coincidence cell. A slightly denser packed hexagonal lattice (a2 = 435 f 10 pm) with a (3 x 3) coincidence cell is found only on a small fraction of the surface. An inherent Moire pattern which can be observed in both cases allows distinguishing the two very similar adsorbate lattices and evaluating adlayer models. Self-assembling of alkanethiols (H-(CH2),SH; n = 1,2,4, 6 , 8,9, 10, 11)and dimethyl disulfide from organic solution results in a hexagonal adsorbate lattice (a3 = 440 % 15 pm). A ( d 7 ~ d 7 ) R 1 0 . 9 ~ coincidencecell, as observed for the sulfide adlayer can be proven only for adsorbed methanethiolate. For an increasing alkyl chain length a monotonousincrease in the tunneling resistance is necessary to maintain molecular resolution. This can be regarded as a direct proof for the presence of the alkyl chains in the tunneling gap during imaging. In the case of hexanethiol the formation of the self-assembled monolayer has also been studied in situ.

h)-

Introduction Interest in structures and properties of self-assembled monolayers (SAMs) a t the metal-liquid interface has grown enormously in recent years. Of these monolayers, those formed on Au(ll1) from thiols-especially alkanethiols-have been the most extensively studied. Their characterization was carried out by a large number of surface-sensitive techniques like FTIR spectroscopy,ls2 GIXD,43LEHD,' TED,7XPS,899SHG1'sl' FTMS,12ellipsometry,13and wetting.14-16 Moreover STM17-25 +

Max-Planck-Institut fur Polymerforschung.

* Johannes-Gutenberg-Universitat.New address: Institut fiir

Physik, Humboldt-Universittit zuBerlin, Invalidenstr. 110,D-10115 Berlin, Germany. Abstract published in Advance ACS Abstracts, December 1, 1994. (1)Porter, M. D.; Bright, T. B.; Allara, D. L.; Chidsey, C. E. D. J.Am. Chem. SOC.1987,109,3559. (2)Nuzzo, R. G.; Dubois, L. H.; Allara, D. L. J . Am. Chem. SOC.1990, 112,558. (3)Hahner, G.; W611, Ch.; Buck, M.; Grunze, M. Langmuir 1993,9, 1955. (4)Fenter, P.; Eisenberger, P.; Liang, K. S. Phys. Rev. Lett. 1993,70, 2447. ( 5 ) Camillone, N., 111;Chidsay, C.; Eisenberger, P.; Fenter, P.; Li, J.; Liang, K.; Liu, G.; Scoles, G. J . Chem. Phys. 1993,99,744. (6)Chidssey, C.; Liu, G.; Rowntree, P.; Scoles, G. J. J . Chem. Phys. 1989,91,4421. (7)Strong, L.; Whitesides, G. M. Langmuir 1988,4,546. (8) Bain, C. D.; Biebuyck, H. A.; Whitesides, G. M. Langmuir 1989, @

00 ac- , 0(&a.

(9) Nuzzo, R. G.; Zegarski, B. R.; Dubois, L. H. J . Am. Chem. SOC. 1987,109,733. (10)Buck, M.; Grunze, M.; Eisert, F.; Fischer, J.; Trager, F. J . Vac. Sci. Technol.,A 1992,10,926. (11)Buck, M.; Eisert, F.; Fischer, J.; Grunze, M.; Trager, F. Appl. Phys. A 1991.53.551. 112)Li, Y.;' Huang, Y. L.; McIver, R. T.; Hemminger, J. C. J . Am. Chem. SOC.1992,114,2428. (13)Porter, M. D.; Bright, T. B.; Allara, D. L.; Chidsey, C. E. D. J . Am. Chem. SOC.1987,109,3559. (14)Bain, C . D.; Troughton, E. B.; Tao, Y. T.; Evall, J.; Whitesides, G. M.; Nuzzo, R. G. J . Am. Chem. SOC.1989,111,321. (15)Bain, C. D.; Whitesides, G. M. J . Am. Chem. SOC.1988,110, 5897. (16)Bain, C. D.; Whitesides, G. M. J . Am. Chem. SOC.1988,110, 6561. (17)Schonenberger,C.; Sondag-Huethorst,J.; Jorritsma, J.; Fokkink, L. Langmuir, 1994,10,3.

and SFM,26$27 as true local probes were used to investigate the metal-adsorbate and adsorbate-air interfaces and confirmed a simple commensurate (43x d3)R3O0 structure, which was found by diffraction experiments to have an adsorbate lattice with a = 500 pm. For this packing a tilt of the alkyl chains of approximately 30"ensures an optimized interchain distance.28 The size of defect-free well-ordered areas can be enlarged by healing out corrosion defects at 350 K.25 Although thiols readily form S A M s on Ag(lll)29only very little work has been performed on this substrate so far. By FTIR spectroscopy3' and Raman spe~troscopy3l~~~ significantly smaller tilts than on Au(111)were reported. In ultrahigh vacuum, dosing of dimethyl disulfide on annealed Ag(ll1) results in a LEED pattern, which can be indexed as two domains of(& x J7)R10.9" coincidence structure.33 It is explained by the cleavage of the S-S bond to form a methanethiolate film. The same structure was found previously for the adsorption of H2S and sulfur onAg(ll1). It was attributed to two-dimensional crystals (18)McCarley, R. L.; Kim, Y. T.; Bard, A. J. J . Phys. Chem. 1993, 97,211. (19)Kim, Y.T.; McCarley, R. L.; Bard, A. J. J . Phys. Chem. 1992, 96,7416. (20)Widrig, C. A.;Alves, C. A.; Porter, M. D. J . Am. Chem. SOC.1992, 113.2805. (2l)Alves, C. A.; Smith, E. L.; Widrig, C. A.; Porter, M. D. Proc. SPIE-Int. Opt. Eng. 1992,No. 1663,125. (22)Edinger, K.; Gdzhauser, A.; Demota, K.; Woll, C.; Grunze, M. Langmuir lM3,9,4. (23)Ross, C. B.; Sun, L.; Crooks, R. M. Langmuir 1993,9,632. (24)Kim, Y.T.;Bard, A. J. Langmuir 1992,8, 1096. (25)Bucher, J. P.; Santesson, L. Kern, K. Langmuir 1994,4,981. (26)Pan, J.; Tao, N.; Lindsay, S. M. Langmuir 1993,9,1556. (27)Butt, H. J.; Seifert, K.; Bamberg, E. J . Phys. Chem. 1993,97, 7316. (28)Sellers, H.; Ulman, A.; Shidman, Y.; Eilers, J. E. J . Am. Chem. SOC.1993,115,9389. (29)Laibinis, P. E.; Whitesides, G. M.; Allara, D. L.; Tao, Y. T.; Parikh, A. N.;Nuzzo, R. G. J . Am. Chem. SOC.1991,113,7152. (30)Walczak, M. M.; Chung, C.; Stole, S. M.; Widrig, C. A.;Porter, M. D. J . A m . Chem. SOC.1991,113,2370. (31)Bryant, M. A,; Pemberton, J. E. J . Am. Chem. SOC.1991,113, 3629. (32)Nemetz, A.;Fischer, T.; Ulman, A.; Knoll, W. J . Chem. Phys. 1993,98,5912. (33)Harris, A. L.; Rothberg, L.; Dubois, L. H.; Levinos, N. J.; Dhar, L. Phys. Rev. Lett. 1990,64,2086. Harris, A. L.; Rothberg, L.; Dhar, L.; Dubois, L. H.; Levinos, N. J. J . Chem. Phys. 1991,94,2438.

0743-7463/95/2411-0506$09.00/0 0 1995 American Chemical Society

Langmuir, Vol. 11, No. 2, 1995 507

Sulfide Adlayers on Ag(ll1)

of y-Ag2S.34-36 For octadecanethiolate, however, Eisenberger et al. proved a larger adsorbate lattice using LEHD and GIXD.37 Recently corrosion of the Ag( 111)surface in the presence of hexanethiol in ethanol was imaged by STM, although the adlayer was not molecularlyresolved.25 Since thiolates on Ag( 111) were never systematically studied by STM and the structure of alkanethiolates (H(CH2),S-) for 1 -e n < 18 is not well known, we have investigated a series of alkanethiolates ( n = 1 , 2 , 4 , 6 , 8 , 9,10, 11)on Ag(ll1). In the following we will start with the discussion of sulfides on Ag( 111)and compare these results with alkanethiolates characterized ex situ at the air-metal or in situ at the liquid-metal interface.

a

+ I

b

H

2 nm

Experimental Section Approximately 200 nm thick Ag(ll1) films (Balzers 99.99%) were vacuum-evaporated (pressure less than 5 x mbar) on freshly cleaved mica at 250-300 "C. This procedure produces (111)-orientedsurfaces with atomically flat terraces as large as several 1000 nm2.38After the evaporation process the samples were immediately placed in diluted aqueous solutions (Millipore) of Na2S and H2S for several minutes. For chemisorption of alkanethiols (H-(CH&SH; n = 1,2,4,6,8,9,10,11)and dimethyl disulfide (Aldrich, used as received) the silver surfaces were exposed to a 1-3 mM solution in ethanol (absolute) or dry n-hexanefor 6-20 h. The modified substrates were then removed and rinsed with the pure solvent and dried in argon gas. H2S and CH3SH were also chemisorbed by gas-phase reaction of the Ag(ll1) surface in a mixture of 90% argon and 10% reactant under atmospheric pressure. The samples were characterized ex situ by a homebuilt STM39under atmospheric conditions. Moreover self-assemblingof hexanethiol was studied in situ from 10to 100mM solution in phenyloctaneon freshly prepared silver. The tunneling tips were electrochemically etched (KOH 2n NaCN 6n) from P t J r (80:20)wire. The images were obtained in the constant height mode under ambient conditions (scan rate, 0.3 Hdimage; tip bias positive). They were stored on a videotape and are presented, if not differently mentioned, unfiltered and unprocessed. Bright areas represent high tunneling currents. The scanner was calibrated for each tunneling tip against graphite. The lattice constants of hexagonal adlayers were determined by averaging over the three main axes on several independent images. This allows reducing the error of the absolute determination of the lattice constants to the range of f 1 0 to f 2 0 pm, depending on the absolute value of the lattice constant. If, in addition, Moire patterns are evaluated, the error can be reduced further.

2 nm

+

C

Results 1. Na2S and H2S. Untreated Ag(ll1) surfaces have been imaged by STM at ambient conditionsand at constant current. They reveal (111)oriented surfaces with atomically flat terraces, as reported earlier.38 After chemisorption of sulfides from aqueous solution, single crystalline adlayer domains with diameters on the order of 10 nm occur. Constant height images on these atomically flat areas exhibit atomic scale resolution. They reveal a hexagonal adsorbate lattice with a next nearest neighbour spacing of a1 = 440 f 10 pm (structure a) (Figure la). Simultaneously a second larger structure ( b l = 760 f 20 pm) can be imaged (Figure lb,c). Figure Id shows both structures superimposed. The larger structure is at(34)Schwaha, K; Spencer, N. D.; Lambert, R. M. Surf.Sci. 1979,81, 273. (35)Rovida, G.;Pratesi, F. Surf. Sci. 1981,104,60. (36)Rovida, G.;Pratesi, F. Proc. Int. Conf.Solid Surf. 1980,321. (37)Fenter, P.; Eisenberger, P.; Li, J.;Cammillone, N., 111; Bernasek, S.; Scoles, G.; Ramanarayanan, T. A.; Liang, K. S. Langmuir 1989,7 , 2013. (38)Buchholz, S.;Fuchs, H.; Rabe, J. P. J. Vac. Sci. Technol. B , 1991,9,857. (39)Rabe, J.P.; Sano, M.; Batchelder, D.; Kalatchev, A. A. J.Microsc. (Oxford)1988,152,573.

d

4-*

2 nm

H 2 nm

Figure 1. Sulfide adlayer (structure a)on Ag(111)from aqueous solution of Na2S (Ut= 400 mV, It= 3 nA): (a)adsorbate lattice (a1= 440 f10 pm); (b)Moir6 pattern between the sulfide adlayer and theAg(ll1) substrate (bl = 760 f 20 pm); (c) simultaneous imaging of a1 and bl; (d)imaging of the sulfide adlayer with the superimposed Moire pattern (band pass filtered).

508 Langmuir, Vol. 11, No. 2, 1995

Heinz and Rube

a

2 nm

H 2 nm

Figure 2. Sulfide adlayer (structure b) on Ag(ll1) from gasphase deposition with H2S (a2= 435 f 10 pm) with an inherent (2 x 2) Moir6 pattern (b2 = 870 f 20 pm) Ut = 300 mV, It = 2 nA, band pass filtered).

b

H 2 nm

H 1 nm Figure 3. Untreated Ag(ll1) surface (a = 285 f 10 pm, Ut = 50 mV, It = 2.5 nA).

tributed to a ( d 3x d3)R3Oo Moire pattern between structure a and the Ag(ll1) lattice. The same kind of images can be obtained from samples treated with H2S in a gas-phase reaction. A second slightly more compressed sulfide adlayer (structure b, a2 = 435 f 10 pm) with a superimposed (2 x 2) Moire pattern (b2 = 870 f 20 pm) is observed only on a small fraction of the sample (Figure 2). The local defects shown in Figure 1are not necessarily intrinsic properties of the adsorbate. They can result from bulk defects and corrosion of the substrate. Atomically resolved images of the Ag(ll1) substrate under the adsorbed sulfide could not be obtained. However on untreated surfaces the Ag(111) lattice is occasionally observed (Figure 3). 2. Alkanethiols and Dimethyl Disulfide. Ex situ imaging at the air-metal interface indicates that the selfassembly of alkanethiols (H-(CH,),SH; n = 1 , 2 , 4 , 6 , 8 , 9, 10, 11)and dimethyl disulfide results in a hexagonal adsorbate structure (a3 = 440 f 15 pm, Figure 4). The same structure was found in the in situ experiment after an exposure of a freshly prepared Ag(ll1) surface to a solution of hexanethiol for several minutes (image not

C

t-----l 2nm

Figure 4. Alkanethiolates on Ag(ll1) from organic solution (a3= 440 f 15pm): (a) ethanethiolate (Ut = 160 mV, I = 6 nA); (b) hexanethiolate (Ut =500 mV,It = 2 nA); (c) undecanethiolate (Ut = 1.5 V, It = 200 PA, band pass filtered).

displayed). Moreover for methanethiol and dimethyl disulfide adsorbed on Ag(111) a ( d 3x d3)R3Oo Moire pattern with bs = 760 f 20 pm can be imaged which has never been observed for any other thiols (Figure 5). For highly resolved images we define a minimal tunneling resistance to maintain molecular resolution (Rmin). It increases monotonously with the number of C atoms in the alkyl chain. Ethanethiolate can be resolved at Rmin

Sulfide Adlayers on Ag(ll1)

Langmuir, Vol. 11, No. 2, 1995 509

Figure 6. Error circle for the construction of the model for the sulfide adlayer; the two circles represent the error for the Moire pattern: 0,silver atoms; 0,equivalent adsorbate sites; -, lattice vectors of the coincidence cell.

H 2 nm Figure 5. Methanethiolate on Ag(ll1)from gas phase; Moire pattern with b~ = 760 f 20 pm (Ut= 150 mV, It = 3 nA). = 10-100 MQ (Ut= 0.1-0.2 V, It < 10 nA),hexanethiolate at R m i n = 0.1-1 GQ (Ut = 0.3-0.6 V, It < 4 nA), and undecanethiolate at R m j n = 10-100 GQ (Ut = 1-1.5 V, It < 200 PA), respectively. It should be mentioned that a variation of R m j n from tip to tip has been observed and that the values given here are the smallest we found. For certain tips these conditions can already be destructive.

Discussion 1. Na2S and H2S. An analysis of the Moire pattern between the sulfide adsorbate and the silver substrate lattice is used to improve the determination of adlayer lattice constants measured by STM and to construct welldefined adlayer models from STM images, respectively. Moir6 patterns are caused by the adsorption to inequivalent binding sites on the substrate. Since only equivalent binding sites can result in locally identical tunneling currents, the lattice constant of a Moire pattern corresponds to very similar or identical positions of the adsorbate on the substrate. If the Moire pattern is commensurate to the adsorbate, it will be commensurate to the substrate too. Consequently it can be identified with the coincidence lattice of the adsorbate which exhibits equal binding sites. Although for proper analysis imaging of the Moire pattern and the adsorbate lattice is necessary, in most cases a STM image is dominated either by the Moire pattern or by the adsorbate lattice, depending on the tip conditions. In Figure Id it is shown that the Moir6 pattern is commensurate for more than 12 unit cells with the sulfide adsorbate (see arrows). Evaluating a large number of similar images leads to the conclusion that the Moire pattern is exactly commensurate with the adsorbate. Therefore V3 of the adsorbed species are localized on equivalent coincidence positions. From this one can deduce a model of structure a which assumes a twodimensional adsorbate layer on a hexagonal nondistorted Ag(111) surface. Figure 6 displays the absolute error of the unit cell of the Moire pattern (bl = 760 f20 pm). The error circle intersects with 12 equivalent binding sites with the same distance from the origin. Assuming a hexagonal symmetry of the adsorbate and the substrate the unequivocal solution gives two domain orientations of the adsorbate. It produces the definite lattice constant and the rotation of the coincidence lattice on Ag( 111).With the assumption that the substrate is not distorted, one finds exactly a1 = 441 pm and bl = 764 pm. The next

Figure 7. Model of the sulfide adlayer (a1= 441 pm, b11 = 764 pm; structure a): smaller circles, silver atoms; larger circles, sulfur atoms; hatched circles, coincidence lattice.

solutions b2 = 867 pm (+13%)and b3 = 578 pm (-32%) are not within the error of the STM measurements. From this a ( 4 7x 47)RlO.g"model for the coincidence cell of the adsorbate can be deduced (Figure 7). In this model the unit cell consists of three sites. One-third of the adsorbed sulfide is bound on top sites, 2/3 on hollowsites (hpc and fcc binding sites). In their ab initio calculations Ulman et al. reported chemically different t and h positions on Ag(ll1) for the adsorption of HS species28. For the t-site adsorption they calculated 3.3 kJ*mol-l less Ag-S binding energy and an approximately 16%larger Ag-S distance. Therefore we exclude a second arrangement of the adsorbed sulfide with an inverted number oft- and h-binding sites. The elevated positions of the sulfide on t-sites may be responsible for the higher tunneling currents in the STM images at the coincidence sites (see Figure Id). In LEED experiments Lambert34 and P r a t e ~ iob~~?~~ served on A 111) surfaces dosed with S:! and H2S the same ( 4 7 x 7)R10.9' structure ( a = 441 pm) which was constructed for the sulfide adlayer in our experiment. Therefore the assumption of a nondistorted underlying Ag(ll1) lattice used in the construction of the adlayer model was justified. From the STM images one cannot directly determine the chemical nature of the sulfide adlayer but it should be mentioned that in the LEED experiments the sulfide adlayer is explained in terms of an epitaxially grown monolayer of two-dimensional (111)oriented y-Ag2S (a = 441 pm) consisting of sulfide and silver ions. Figure 8 displays the model of the hexagonal structure b (a2 = 434 pm, b2 = 867 pm), which is 1.6% more compressed than structure a. The model has been constructed by an error circle similar to Figure 6 and results in an unrotated (3 x 3) coincidence structure. The t sites for the coincidence lattice are arbitrarily chosen. By comparison of the Moire patterns of structures a and b both structures can easily be distinguished, although

f

510 Langmuir, Vol. 11, No. 2, 1995

Figure 8. Model of the slightly compressed sulfide adlayers (a2= 434 pm, bz = 867 pm; structure b); smaller circles, silver atoms; larger circles, sulfur atoms;hatched circles,coincidence lattice. their lattice constants (al, az) are the same within the error of the absolute adsorbate distance measurement in STM. 2. Alkanethiols and Dimethyl Disulfide. Atomic scale resolution images of untreated Ag(ll1) surfaces in air can only occasionally be observed. This is attributed to a poorly ordered adlayer containing oxygen, carbon, and possibly traces of chlorine and sulfur, as detected by The treatment with the thiol solutions replaces this layer. This is evidenced by XPS studies on octadecanethiol on Ag(ll1) by Laibinis which show that the residual oxide content of the adlayer on silver is very low, perhaps of the order of a few percent of a monolayer. It was concluded that the adsorption of thiols on silver reverses oxidation ofthe substrate that may have occurred and displaces any adsorbed oxygen species.29 Sulfides are assumed to behave similarly. Chemisorption of alkanethiols on Ag(ll1) is supposed to follow a redox process on the substrate, which results in thiolate monolayers after previous cleavage of the S-H bond similar to the case of Au(111).12 From the in situ STM imaging of the adsorption of hexanethiol at the liquid-metal interface, one can conclude that in 10-100 mM solution of hexanethiol the self-assembly on Ag(ll1) occurs on a time scale of a few minutes. This is believed to be similar for any other short alkanethiols but dependent on the thiol concentration in the solution and on the contamination of the silver surface. On Au(ll1) the formation of highly ordered S A M s of long chain thiols from very diluted solution occurs significantly slower. Buck et al. studied the formation of SAMs of long chain thiols with XPS,loJ1 SHG,l0J1 and In very diluted solutions (micromoles per liter range) they proved a two-step mechanism: The chemisorption of thiols on Ag(ll1) was observed after several minutes. However, the formation of highly ordered SAMs took several hours. The difference in ordering kinetics can possibly be explained by the significantly shorter alkyl chain length in our experiment. Alternatively the ordering kinetics may be different for perpendicularly oriented thiols on Ag(ll1) or tilted thiols on Au(ll1). For the formationsof S A M s on Ag(111)the driving force, especially for short chain homologues, is the binding energy of the sulfur headgroup to the Ag( 111)substrate.28 Consequently the question arises whether the alkyl chains can adopt the packing discussed for the sulfide adlayer above (a1 = 441 pm). For adsorption of sulfides, methanethiol, and dimethyl disulfide, the same adsorbate lattice and Moire pattern are observed. Therefore we conclude that methanethiolate can adopt the sulfide adlayer lattice with a ( 4 7x 47)R10.9' coincidencecell and an area per molecule of 16.87 k . This structure is (40) Sagiv, J., Maoz, R. Unpublished results.

(41)Kitaigorodskij,A. I.; Mnjukh, J. B. Bull. Acad. Sci. USSR,Diu. C h e n . Sci. (Engl. Transl.) 1959, 1992.

Heinz and Rube confirmed by LEED investigations where gas-phase adsorption of dimethyl disulfide leads to a monolayer of methanethi~late.~~ From the presence of the CH3 group which is vertically oriented with respect to the Ag(ll1) surface, Harris et al. excluded the formation of a sulfide adlayer by cleavage of the S-C bond.33 For thiolates with n > 1, Moire patterns were never observed. Therefore one can exclude imaging of simple sulfide adlayers formed by decomposition of the thiols. The construction of an adlayer model is impeded by the lack of Moire pattern. Electronic and stuctural effects have to be considered for an explanation: The increasing length of the hydrocarbon chain should change the electronic properties of the sulfur headgroup (e.g. charge density). Ulman et al. calculated for adsorbed CH3S that the t and h binding sites are more similar than for adsorbed HS.28 The difference in bond length decreased from 16% to 13.5%. Therefore the observation of a Moire pattern is more unlikely with an increasing number of methylene units. However, the electronic properties are most strongly influenced by the next bound atom which changes from HS to CH3S but not from methanethiolate to ethanethiolate, where the disappearance of the Moire pattern was observed in our experiments. Therefore we do not believe that this effect is sufficient to explain the lack of Moire patterns for n > 1. On the other hand the absence of the Moire pattern can be better explained by a loss of commensurability between the thiolate and the Ag(ll1) substrate. Alkanethiolates with n > 1 exhibit close packing of perpendicularly oriented alkyl chains as deduced from a chain tilt angle of -10°.29,31They have an elliptical van der Waals cross section parallel to the surface, which is larger than for methanethiolate. From comparison with X-ray data of the triclinic octadecane the two main axes of the cross section should be about 420 and 480 pm, re~pectively.~~ Therefore the cross section would be considerably compressed if the adlayer lattice of methanethiolate (a = 441 pm) were adopted. Consequently the adsorbate lattice constant of thiolates with n > 1 could be slightly enlarged within the error of the experiment (a3= 440 f15pm), possibly to reduce repulsive interactions which should increase with the number of methylene units in the alkyl chain. An increase of the lattice constants of thiolates with an increasing chain length has also be considered because Eisenberger et al. observed an incommensurate adlayer (a = 477 f 3 pm) for octadecanethiolate on Ag(ll1) by GMD.37 On Au(ll1) thiolates are more than 20% less dense packed (a = 500 pm). STM images of this structure have been attributed to a commensurate ( 4 3x d3)R3Oostructure of the thiolates and were regarded as images of the sulfur headgroup.22 However Bard et al. could not interpret their investigations on structurally demanding thiols by this explanation of the contrast.lg Curiously Porter et a1.20could image short chain and long chain thiols using similar low tunneling resistances at tunneling currents around 2 nA, although Bard et al. could not reproduce their experiments.18 The latter imaged c18H37S films at a tunneling resistance of roughly 150 MQ and observed similar corrosion patterns as for NazS. They interpreted them as tip-induced etching of the organic film. Recently Schonenberger et al. obtained atomically resolved images for C12H25S films at significantly higher tunneling resistances. l7 In our experiments on Ag(lll), thiolates with longer alkyl chains could not be molecularly resolved with the same minimal tunneling resistance (Rmln) as the thiolates with shorter alkyl chains. In general, to maintain molecular resolution of the thiolates, Rmin has to be monotonously increased with the number of methylene

Sulfide Adlayers on Ag(ll1)

units in the alkyl chains. The increase amounts to a factor of 10-100 for 5 units. This is regarded as a direct proof for the presence of alkyl chains in the tunneling gap. Since an increasing tunneling resistance results in an increasing tip-surface distance, we conclude further that a minimal tip-surface distance is necessary to ensure nondestructive imaging of thiolates depending on the length of the alkyl chain.

Conclusions Two very similar sulfide adlayers on Ag(ll1) can be distinguished by analyzing Moire patterns between adsorbate and substrate in STM images. From ex situ imaging of adlayer lattices, inherent Moire patterns, and the construction of adlayer models, one can conclude that methanethiolate adopts the hexagonal packin of the larger sulfide adlayer (a1 = 440 f 10 pm) with a ( $ 7 ~47)R10.9” coincidence cell. The smaller sulfide adlayer (a2 = 435 f 10 pm) with a (3 x 3) coincidence cell is not adopted. Due to the lack of Moire patterns in images of thiolates with n > 1,a small widening of the adlayer lattice

Langrnuir, Vol. 11, No. 2, 1995 511

within the error of the experiment (a3 = 440 f 15 pm) is possible. To maintain molecular resolution of the thiolates, the minimal tunneling resistance (R,i,J has to be monotonously increased with the number of methylene units in the alkyl chains. The increase amounts to a factor of 10-100 for 5 units. This is regarded as a direct proof for the presence of the alkyl chains in the tunneling gap. Self-assembly of hexanethiolate can also be imaged in situ at the liquid-metal interface and has been shown to occur on a time scale of a few minutes.

Acknowledgment. The authors wish to thank Professor J. Sagiv and Dr. R. Maoz (Weizmann Institute of Science) for very helpful discussions and the Institut fur Mikrotechnik Mainz for support in the preparation of the silver substrates. The project is supported by the German Israeli Foundation and the Bundesministerium fur Forschung and Technologie (“Muster selbstorganisierender Molekule”). R.H. acknowledges a Kefule scholarship granted by the Fonds der Chemischen Industrie. LA940401Q